Method for determining, in real time and continuously, an amount of particles of a given material

Abstract

The invention describes a method for determining, in real time and continuously, an amount of particles of a given material in a sample, comprising steps of: E1: optically measuring a signature of the sample by means of an optical sensor (10); E2: identifying, by means of a processing unit (20), the type of the sample by means of a classification model trained on a training database comprising a plurality of reference signatures, each reference signature being associated with a reference sample type; and E3: determining, by means of the processing unit (20), the amount of particles of the given material in the sample based on the identified sample type and a correspondence table associating, with each of a plurality of reference sample types, a reference amount of particles of the given material.

Description

FIELD OF INVENTION

[0001] The present invention relates to the field of air quality monitoring and concerns a method for determining, in real time and continuously, the particle concentration of a given material in a sample, and a system for determining, in real time and continuously, the particle concentration of a given material in a sample. More particularly, the present invention can be used to determine, in real time and continuously, the concentration of respirable crystalline silica emitted during a construction project. STATE OF THE ART

[0002] Air quality is impacted by the type and number of particles suspended in the air. Air quality can have consequences for the health of people breathing that air, particularly when certain potentially harmful particles are present.

[0003] In particular, on construction sites, air quality is likely to deteriorate due to the emission of dust containing particles of certain materials, which can pose hazards to workers when inhaled for extended periods. For example, respirable crystalline silica dust, ballast dust, lime, other hydraulic binders, contaminated soil dust, chromium VI particles, or diesel particles are likely to be emitted by various construction operations.

[0004] Certain types of particles, such as respirable crystalline silica, pose a particularly high health risk, with adverse effects. Crystalline silica is classified as respirable when inhaled particles have a diameter of less than 10 µm, allowing them to reach the alveoli of the lungs and cause pathologies such as silicosis or cancer. Work involving exposure to respirable crystalline silica dust is classified as a carcinogenic activity as of January 1, 2021.

[0005] It is therefore important to monitor air quality, that is to say to identify and determine the rate of particles of a given material, in order to put in place protective measures for operators working on construction sites which are adapted according to the type and quantity of particles emitted.

[0006] Current air quality monitoring methods involve taking spot samples at a construction site. These samples are then tested in the laboratory, their chemical analysis allowing the determination of the level of particles of given materials, for example respirable crystalline silica, in the sample.

[0007] The collection and analysis of samples is a complex and lengthy process, requiring at least a full workday. A device is worn by an operator throughout the day, accumulating particles, and is then analyzed the following day in a laboratory.

[0008] Consequently, current methods do not allow for risk assessment, either continuously or in real time, or for all operations carried out on construction sites, for operators present on the sites. Therefore, operator protection and alert measures cannot be adapted to the actual, real-time, continuous particle levels. As a result, operators can be exposed to levels of particles of potentially hazardous materials, such as respirable crystalline silica, ballast dust, etc., for a relatively long period corresponding to the time between two samplings. The management of health risks related to operator exposure to particles of certain types of materials is therefore inadequate.

[0009] US patent 2015 / 068806 A1 describes a method for the non-intrusive analysis of compounds present in drilling debris. US patent 2016 / 041074 A1 describes a method for analyzing and filtering air using filtration units that adapt the air filtration according to the analysis. DESCRIPTION OF THE INVENTION

[0010] An objective of the invention is to propose a method for determining the particle level of a given material in a sample that is more efficient than current methods, in particular that allows for real-time and continuous determination of the particle level of the given material in an air sample.

[0011] To this end, the invention describes a method for determining the rate of particles of a given material in an air sample according to independent claim 1.

[0012] Some preferred but not limiting characteristics of the method for determining the particle content of a given material described above are the following, taken individually or in combination: the given material is at least one of the following: alveolar crystalline silica, ballast dust, lime, diesel particles, contaminated soil dust, other hydraulic binders, Chromium VI particles, and / or other particles emitted by various construction site operations; the signature of the sample measured in step E1 is a function of the size and number of particles present in the sample; the signature of the sample measured in step E1 includes a histogram of the distribution of the number of particles in the sample as a function of particle size; the type of sample identified in step E2 is defined by the nature of the sample and the activity of obtaining the sample;The signature of the sample is measured periodically in step E1 at each measurement period; the type of sample is identified periodically in step E2 at each calculation period; and the particle rate is determined periodically in step E3 at each calculation period. The calculation period is equal to or greater than the measurement period, and the measurement and calculation periods are between 1 second and 1 hour. The process further includes a step E4 in which the processing unit determines exposure to particles of the given material based on the particle rate of the given material determined in step E2 and the exposure time to said particles of the given material. The process further includes a step E5 in which a user is alerted when exposure to particles of the given material exceeds a predetermined exposure threshold.

[0013] According to a second aspect, the invention also describes a system for determining the rate of particles of a given material in an air sample according to independent claim 10.

[0014] The optical counter can be portable.

[0015] The system for determining the rate of particles of a given material in a sample described above may further include storage means configured to store the measurements made by the optical counter and / or the training database and / or the lookup table. DESCRIPTION OF THE FIGURES

[0016] Other features, purposes and advantages of the present invention will become apparent from the detailed description that follows, given by way of non-limiting example, which will be illustrated by the following figures: There figure 1is a block diagram representing different stages of a process for determining a particle count according to an embodiment of the invention. figure 2 is a block diagram representing different stages of a process for determining a particle count according to an embodiment of the invention. figure 3 is a block diagram representing different stages of a process for determining a particle count according to an embodiment of the invention. figure 4 is an example of a histogram showing the distribution of the number of particles as a function of their size. figure 5 is a diagram representing different elements of a system for determining a particle rate according to an embodiment of the invention. DETAILED DESCRIPTION OF THE INVENTION

[0017] A method for determining, in real time and continuously, the particle concentration of a given material in a sample, as illustrated by way of non-limiting example in figure 1, in figure 2 and in figure 3 , includes the following steps: E1: optical measurement of a signature of the sample; E2: identification by a processing unit 20 of a type of the sample by means of a classification model trained on a training database comprising a plurality of reference signatures, each reference signature being associated with a type of a reference sample; and E3: determination by the processing unit 20 of the rate of particles of the given material in the sample from the type of the identified sample and a lookup table associating to each of a plurality of types of reference samples a rate of reference particles of the given material.

[0018] The process makes it possible to determine the rate of particles of a given material present in a sample, that is to say, to differentiate, within the particles present in the sample, the particles of the given material from the particles of other materials.

[0019] By "sample", we will understand that it is a sample of air likely to include suspended particles, for example dust particles or particles of given materials, and which may in particular be breathed by an individual working on a construction site.

[0020] Of course, several particle levels of several different given materials can be determined in step E3. Thus, particles of one or more materials potentially harmful to health can be discriminated against particles of other materials that do not present a health hazard.

[0021] The process can be implemented, for example, on construction sites, to improve the management of health risks for individuals working on the sites, such as operators or inspectors.

[0022] The process allows for the determination of the particle concentration of a given material in the sample in real time and continuously. "Real time" means that the determination is performed simultaneously with the individual's exposure to the particles of the given material, for example, during the operator's or inspector's work on the construction site. "Continuous" means that the determination is performed regularly over time, that is, periodically, and at intervals that allow for the rapid adaptation of protective measures based on the determined particle concentration of the given material, and therefore on the individual's exposure to the particles of the given material. For example, the particle concentration can be determined throughout an entire workday, with a calculation period ranging from 1 second to 1 hour.

[0023] Thus, the risk resulting from an individual's exposure to particles of a given material can be assessed in real time and continuously. Protective and alert measures for the individual can therefore be adapted in real time and continuously, preventing excessive exposure to materials potentially harmful to their health. The management of health risks related to individual exposure to particles of certain materials is therefore improved.

[0024] The sample type is determined using a classification model. This allows for reliable identification of the sample type, thanks to the classification model being trained on the training dataset.

[0025] Furthermore, sample type assignment can be performed even in complex cases where the sample does not correspond to a single reference sample type, but rather to a combination of several reference sample types, for example, in the case of dust mixtures. The classification model thus allows for the isolation of the different types of dust detected. Therefore, based on the sample signature measured in step E1, the sample type identification performed in step E2 leads to the assignment of a sample type that may correspond either to a single reference type or to a combination of several reference types.

[0026] The sample material can be respirable crystalline silica. This allows for the real-time and continuous determination of the respirable crystalline silica level emitted during construction operations. Alternatively, or in addition, the sample material can be ballast dust, lime, other hydraulic binders, contaminated soil dust or other types of dust, or chromium VI particles. , or diesel particles, and / or other particles emitted by various construction site operations. Step E1: Optical measurement of a sample signature

[0027] Optical measurement of the sample's signature is performed in real time and continuously at stage E1. Thus, the measurement data reflects the actual air quality breathed by the individual, particularly the operator during their work on the construction site, and the evolution of this air quality over time.

[0028] The sample signature can be measured periodically at step E1 during each measurement period. This measurement period can be between 0.1 seconds and 10 hours, for example, between 1 second and 1 hour, for example, between 1 minute and 30 minutes, for example, 15 minutes. In the latter case, the sample signature is measured every 15 minutes.

[0029] The measurement of the sample signature is carried out using an optical sensor 10 which is an optical counter configured to perform particle counting.

[0030] The signature of the sample measured in step E1 is a function of the size and number of particles present in the sample. The signature of the sample therefore depends on the size and number of particles present in the sample.

[0031] The signature of the sample measured in step E1 then includes a histogram of the distribution of a number of particles in the sample as a function of the size of the particles, that is to say a histogram of the size distribution of an alveolar fraction of particles of the sample.

[0032] The concentration of particles of a given material in the sample, expressed for example in ppm or µg / m³, can be deduced from the histogram showing the distribution of the number of particles as a function of their size. For example, the concentration can be determined by assuming that all particles are spherical and by considering known material density values.

[0033] The signature of the sample, in particular the histogram of the distribution of the number of particles in the sample as a function of their size, can be measured in a particle size range that corresponds to particles smaller than 40 µm, for example sizes between 0.1 and 10 µm.

[0034] The size range can be divided into several bands, with the number of particles measured in each band. Each band can correspond to a size interval of approximately 1 µm. The size range can be divided into a number of bands between 2 and 40, for example, 24 bands.

[0035] There figure 4This illustrates an example of a histogram showing the distribution of the number of particles as a function of their size. Measurements were taken for a particle size range of 0.1 to 2 µm, divided into 6 slices, and for graphite and silica particles. The distribution of graphite particles as a function of their size is shown in black, and the distribution of silica particles as a function of their size is shown in gray.

[0036] Alternatively, or in addition, the sample signature measured in step E1 is a function of particle shape and particle number. Particles of a given material are then distinguished from particles of other materials present in the sample based on their shape, for example, spheroid, lenticular, cubic, flat, disc, cylindrical, etc. The sample signature thus comprises a distribution of the number of particles in the sample according to their shape.

[0037] Furthermore, the sample signature measured in step E1 can be a function of the number of particles present in the sample and the dimensions of the particles in one direction (e.g., the direction of the particle's largest dimension), two directions, or three orthogonal directions forming a three-dimensional space. The dimensions of a particle in three directions are its width, depth, and height.

[0038] The shape of the particles of the materials present in the sample can be determined based on the dimension(s) of the particles measured in the sample's signature. For example, a spherical or cubic shape of a particle can be identified when the three dimensions of the particle are approximately equal, that is, when the particle has the same width, depth, and height.

[0039] A flat or disc shape of a particle can be identified when the height of the particle is much less than the width and depth of the particle, the width and depth of the particle being otherwise substantially equal or the width not measuring more than twice the depth of the particle, for example when the height of the particle is less than the width and depth of the particle by at least a factor of two or at least a factor of ten.

[0040] A cylindrical disk shape of a particle can be identified when the height of the particle is much greater than the width and depth of the particle, the width and depth of the particle being otherwise substantially equal or the width not measuring more than twice the depth of the particle, for example when the height of the particle is greater than the width and depth of the particle by at least a factor of two or at least a factor of ten.

[0041] Alternatively, or in addition, the sample signature measured in step E1 can be a function of the number of particles present in the sample and the color, i.e., the wavelength, of the particles. The wavelength of visible light is approximately between 380 nm for violet and 780 nm for red.

[0042] Alternatively, or in addition, the sample signature measured in step E1 can be a function of the number of particles present in the sample and their diffraction properties, for example, the particle scattering intensity and / or the particle scattering angle. The scattering intensity reflects the degree to which light is diffracted in a given direction by a particle illuminated by incident light, and the scattering angle corresponds to the given direction in which the diffraction is measured.

[0043] The table below illustrates a non-limiting example of reference signatures measured for three different aerosols, the reference signatures being a function of particle dimensions and shapes, particle color, particle scattering intensity, and particle scattering angle.

[0044] Measuring a sample signature that depends on the number of particles present in the sample, based on their size, shape, dimensions, color, and / or diffraction properties, allows for particle count determination using a miniature, portable, self-contained device attached to the operator, and by an operator without special qualifications. This simplifies the process and reduces costs. [Table 1] Signature of the reference sample Particle dimensions Particle shapes Diffraction properties of particles Particle color X Y Z Diffusion intensity (%) Diffusion angle (°) Wavelength (nm) Aerosol 1 1 1 1 Spherical / cubic 10 60 250 Aerosol 2 1 5 6 Plate / Disc 10 20 600 Aerosol 3 10 1 1 Cylindrical 10 180 250

[0045] The particle dimensions in the sample shown in the table below are given as relative quantities, therefore without units. A sample particle has a given size that belongs to a size class from bin 00 to bin XX. The size of a particle is an aerodynamic size expressed in micrometers (µm) or nanometers (nm).

[0046] For example : Bin class 01 can correspond to an aerodynamic particle size of approximately 0.5 µm, with X = 0.1 µm, Y = 0.5 µm, and Z = 0.5 µm, which corresponds to a flat or disc shape of the particles; bin class 01 can correspond to an aerodynamic particle size of approximately 0.5 µm, with X = 0.5 µm, Y = 0.5 µm, and Z = 0.5 µm, which corresponds to a spherical or cubic shape of the particles; bin class 02 can correspond to an aerodynamic particle size of approximately 1 µm, with X = 1 µm, Y = 0.3 µm, and Z = 0.2 µm, which corresponds to a cylindrical shape of the particles; The bin 02 class can correspond to an aerodynamic particle size of about 1 µm, with X = 0.2 µm, Y = 1 µm, and Z = 1 µm, which corresponds to a flat or disk shape of the particles; etc.

[0047] In summary, the signature of the sample measured in step E1 is a function of the number of particles present in the sample and: of a size, and / or shape, particles present in the sample, and can also be a function of the number of particles present in the sample and: of a dimension, and / or a color, and / or diffraction properties, of the particles present in the sample. Step E2: Identification of a sample type by a processing unit 20

[0048] In a first embodiment, illustrated by way of non-limiting example in figure 2 , the identification of the sample type during step E2 includes a comparison of the sample signature with a plurality of reference signatures stored in the training database, each reference signature being associated with a reference sample type.

[0049] The training database is consulted each time the sample type is identified.

[0050] To determine the type of sample, the signature of the sample is compared with several, in particular with each, of the reference signatures stored in the training database.

[0051] The classification model is a k-nearest neighbors type model. The k types of reference samples whose reference signatures are closest to the measured sample's signature, within a specified distance, are considered to determine the sample type. The value of k can range from 1 to N, where N is the number of samples in the classification database. For example, the value of k can range from 1 to 10, or for instance, be equal to 5.

[0052] In a first example of implementation, the type of sample determined in step E2 corresponds to the most represented type of reference sample among the k reference samples whose reference signatures are closest to the signature measured in step E1.

[0053] In a second embodiment example, the type of sample determined in step E2 corresponds to a combination of the types of the k reference samples whose reference signatures are closest to the signature measured in step E1, for example by assigning to each reference sample among the k closest reference samples an identical weight, or a variable weight depending on the distance of their respective reference signatures to the signature of the sample measured in step E1.

[0054] The process may include a step of determining an unknown type of sample when all distances between the signatures of the reference samples and the signature measured in step E1 are greater than a certain threshold distance.

[0055] In a second embodiment, the classification model is a neural network type model, support vector machine, stochastic gradient algorithm, or random forest algorithm.

[0056] The process then includes a preliminary step of training the classification model from the training database.

[0057] Once the training has been completed, the determination of the sample type can be carried out for a sample signature measured in step E1 via the classification model developed during training, without requiring consultation of the training database.

[0058] The classification model may include alpha-beta pruning.

[0059] In the first embodiment and in the second embodiment, the training database includes a plurality of reference signatures, each reference signature corresponding to a type of a reference sample, which can be obtained during a specific operation.

[0060] Each reference signature can correspond to a reference histogram of the distribution of a number of particles in the reference sample as a function of particle size and / or particle shape and / or particle dimension and / or particle color and / or particle diffraction properties.

[0061] Each signature of a reference sample can correspond to a vector, in a multidimensional feature space constructed using the distribution histogram in size and / or shape and / or dimension and / or color and / or diffraction properties of the particles in the reference sample, each vector corresponding to a type of the reference sample.

[0062] Information regarding the size, color, and / or diffraction properties of the particles can be obtained independently for each size class bin 00- bin XX of the particles in the sample, or be obtained only for some of the size classes bin 00 - bin XX.

[0063] In summary, the signature of the sample measured in step E1 includes: a histogram showing the distribution of the number of particles in the sample as a function of particle size, and / or a histogram showing the distribution of the number of particles in the sample as a function of particle shape, and may also include: a histogram of the distribution of the number of particles in the sample as a function of the size of the particles, and / or a histogram of the distribution of the number of particles in the sample as a function of the color of the particles, and / or a histogram of the distribution of the number of particles in the sample as a function of the diffraction properties of the particles.

[0064] The signature of the sample measured in step E1 corresponds to an unlabeled vector, which in the first embodiment is compared to the vectors corresponding to the reference signatures of the reference samples in the training database.

[0065] A reference sample type can be defined by the nature of the reference sample and the activity involved in obtaining it. Indeed, the nature and / or the activity involved in obtaining a reference sample impacts its signature; in particular, it can impact the size and number distribution of particles in the reference sample, and / or the percentage of particles of a given material in the reference sample.

[0066] The nature of the reference sample may correspond to a material of the reference sample, such as brick, marble, concrete, quartz, the type of dust, etc., more particularly to a material from which the dust contained in the sample originates.

[0067] The activity of obtaining the reference sample may correspond to an activity following which the reference sample was obtained and / or to a tool corresponding to said activity, for example digging, sanding, cutting, demolition, etc., using a circular saw, a sander, a jackhammer, etc.

[0068] For example, a type of reference sample may be a cement brick obtained by sawing with a circular saw, the resulting dust comprising a given fraction of concrete and marble mixture.

[0069] The type of sample identified in step E2 can be defined by a nature of the sample and an activity of obtaining the sample, the type of sample being determined according to the types (nature and activity of obtaining) of the reference samples.

[0070] The more reference sample types the training database contains, the more reliably and efficiently the classification model will be able to recognize a large number of sample types.

[0071] The correspondence table associates a plurality of reference sample types with respective reference particle rates.

[0072] Each type of reference sample is associated with at least one reference particle level of at least one given material. Each type of reference sample can be associated with various reference particle levels of different given materials that may pose a health risk to an individual who breathes them. For example, each type of reference sample could be associated with a level of respirable crystalline silica, a level of ballast dust, and a level of chromium. VI,a rate of lime, a rate of diesel, a rate of contaminated soil dust, a rate of other hydraulic binders, etc.

[0073] The training database and the lookup table can be a single database. The table below illustrates a non-limiting example of a single database combining a training database that associates a reference signature with a reference sample type, and a lookup table that associates each reference sample type with particle levels of several given materials. [Table 2] Signature of the reference sample Type of reference sample Particle level in the reference sample Acquisition activity Nature Alveolar Si level Chromium VI levels X-ray particle rate S1 Digging A1 a1 b1 c1 S2 Sanding A2 a2 b2 c2 S3 Cutting A3 a3 b3 c3 S4 Demolition A4 a4 b4 c4 ... ... ... ... ... ...

[0074] An example of a method for constructing the training database and the lookup table is described in the following paragraphs.

[0075] To build the training database, a plurality of reference samples with known reference types are collected on site. A reference signature corresponding to each reference sample is established, either directly on site, for example using an optical sensor adapted for particle counting, or subsequently, for example through laboratory analysis.

[0076] In order to construct the correspondence table, particle levels of various given materials corresponding to the type of reference sample are analyzed, for example through a chemical analysis carried out in the laboratory, in order to associate with each type of reference sample the particle levels of the various given materials. Step E3: Determination by processing unit 20 of the particle content of the given material in the sample

[0077] Once the sample type is identified, the processing unit 20 determines the particle rate of a given material in the sample, using the particle rate of the given material associated with each reference sample type.

[0078] The sample type can be identified periodically in step E2 at each calculation period. The particle rate can be determined periodically in step E3 at each calculation period.

[0079] The calculation period can be equal to or greater than the measurement period. The identification of the sample type and / or the determination of the particle content of the given material can therefore be carried out at each measurement performed by the optical sensor 10, or at longer time intervals, so as to monitor in real time and continuously, and therefore regularly over time, the evolution of the particle content of the given material in the sample.

[0080] The calculation period can be equal to the measurement period. The calculation period can be between 0.1 seconds and 10 hours, for example between 1 second and 1 hour, for example between 1 minute and 30 minutes, for example equal to 15 minutes.

[0081] In particular, when the measurement period and the calculation period are both equal to 15 minutes, the signature of the sample is measured every 15 minutes, and the particle rate of the given material is determined every 15 minutes, i.e., at each measurement of a signature of a sample.

[0082] When the calculation period is longer than the measurement period, the signature of the sample used to determine the sample type can be an average of the signatures of the samples obtained at each measurement period occurring during the calculation period. This improves the device's autonomy and smooths out potential measurement errors. E4: Determination of exposure to particles of the given material

[0083] The process may further include a step E4 of determining by the processing unit 20 an exposure to particles of the given material from the particle rate of the given material determined in step E2 and an exposure time to said particles of the given material, as illustrated by way of non-limiting example in figure 3 .

[0084] Exposure to the given material type during a calculation period can be calculated by integrating the particle concentration of the given material determined in step E3 over a duration corresponding to the calculation period. Exposure to the given material type during the exposure time determined in step E4 corresponds to the sum of exposures during the calculation periods corresponding to this exposure time. Thus, the exposure calculated in step E4 takes into account the evolution over time of the particle concentration of the given material determined in step E3.

[0085] The exposure time can be greater than or equal to the calculation period. The exposure time can be defined according to the type and toxicity of the given material, and / or according to the desired speed of adaptation of protective measures. Thus, the shorter the exposure time, the faster and more accurately the exposure determination can be performed.

[0086] The exposure time can be between 0.1 seconds and 10 hours, for example between 1 second and 1 hour, for example between 1 minute and 30 minutes, for example be equal to 15 minutes.

[0087] For example, if the exposure time is equal to 1 hour and the calculation period is equal to 15 minutes, the exposure is determined by summing, for each of the 4 calculation periods included in the exposure time, the particle levels determined at each of these 4 calculation periods.

[0088] Alternatively, or in addition, a predicted exposure to a given type of particle can be determined by multiplying the particle concentration of the given material, as determined during the calculation period, by a predicted exposure time corresponding to several calculation periods. Thus, a predicted exposure to the given type of particle can be calculated in advance of the actual exposure. E5: User alert

[0089] The process may further include a step E5 for alerting a user when exposure to particles of the given material exceeds a predetermined exposure threshold, as illustrated by way of non-limiting example in figure 3 .

[0090] The exposure threshold can be defined according to the given material, its potential impact on the health of the individual likely to breathe it, specific safety standards, etc.

[0091] Alternatively or in addition, the process may include a user alert step E5' when the particle rate determined in step E3 during a calculation period exceeds a predetermined rate threshold.

[0092] Thus, a temporary but significant exceedance of a given material particle level, regardless of when the exceedance occurs and the duration of the exceedance, is identified and the user is alerted, in order to be able to take appropriate protective measures in real time and continuously.

[0093] Alternatively, or in addition, the process may include a step E5" that alerts a user when the predicted exposure to the given type of particle material exceeds a predetermined exposure threshold. This allows protective measures to be taken before the operator is actually exposed to the threshold. E6: Alert data transmission

[0094] The process may further include an alert data transmission step E6. The alert data may include the exposure determined in step E5, E5', or E5" and / or the particle rate determined in step E3 for each calculation period and / or the signatures measured in step E1 for each calculation period.

[0095] Alert data can be transmitted when exposure to particles from the given material exceeds a predetermined exposure threshold, and / or when the particle rate determined in step E3 during a calculation period exceeds a predetermined rate threshold.

[0096] Alert data can be transmitted to a remote server via a wireless connection. System for determining, in real time and continuously, the particle concentration of a given material in a sample

[0097] A system for the real-time and continuous determination of the particle concentration of a given material in a sample includes, as illustrated by way of non-limiting example in figure 5 : an optical sensor 10 adapted to measure a signature of a sample; and a processing unit 20 configured to: identify a type of the sample by means of a classification model trained on a training database comprising a plurality of reference signatures, each reference signature being associated with a type of a reference sample, and determine the rate of particles of the given material in the sample, from the type of the identified sample and a lookup table associating each of a plurality of reference sample types with a rate of reference particles of the given material.

[0098] The system is suitable for implementing the process described above. The system can be used to identify and determine the concentration of respirable crystalline silica particles emitted by various construction site operations. Alternatively, or in addition, the system can be used to identify and determine the concentration of ballast dust, lime, diesel, contaminated soil dust, other hydraulic binders, Chromium VI, and / or other particles emitted by various construction site operations.

[0099] The term "processing unit 20" can refer to any system capable of performing the desired processing, for example, a single processing unit 20, or several processing units 20 that can be controlled by a master unit. The processing unit 20 may include one or more microprocessors adapted to process the measurement data from the optical sensor 10 in order to determine the particle concentration of the given material in the sample.

[0100] The classification model, training database, and lookup table can correspond to those described above concerning the process of determining a particle rate of the given material in a sample.

[0101] The processing unit 20 can further be configured to determine exposure to particles of the given material from the determined particle rate of the given material and a time of exposure to said particles of the given material.

[0102] The optical sensor 10 is an optical particle counter configured to perform particle counting. More specifically, the optical sensor 10 can be configured to generate a distribution histogram of the number of particles in the sample as a function of particle size. A concentration of the particles of a given material in the sample, expressed for example in ppm or µg / m³, can be deduced from the distribution histogram measured by the optical sensor 10. For example, the concentration can be determined by assuming all particles are spherical and considering known material density values.

[0103] The optical sensor 10 can be adapted to accurately measure the signature of the sample, more specifically the distribution of the number of particles according to their size, in a particle size range that can correspond to particles smaller than 40 µm, for example between 0.1 and 10 µm.

[0104] The optical sensor 10 can be a micro-sensor. The optical sensor 10 can be a nephelometer-type optical sensor with laser diffraction. Such optical particle counters are robust, accurate, and inexpensive, and allow for reliable measurements on a wide variety of sample types, whether they come from natural rocks (granite, sand, etc.) or processed materials (concrete, cement, mortar).

[0105] The optical sensor 10 can be portable. By portable, we mean that the optical sensor 10 is not permanently fixed to the ground, can be transported, and has a moderate weight and size. In particular, the optical sensor 10 can be carried by a user, such as an individual working on a construction site, like an operator or inspector, without causing significant disruption to their work.

[0106] The measurements taken by the optical sensor 10 can be transmitted to the processing unit 20, which can be a remote server. Alternatively, the processing unit 20 and the optical sensor 10 can be integrated into a portable module. Exposure can then be assessed on a personal level.

[0107] Alternatively, the optical sensor 10 can be stationary, meaning it can be positioned at a specific location on the construction site. The processing unit 20 can be a remote server to which the measurement data is sent for processing. The system then includes means for transmitting the measurement data to the processing unit 20. Exposure can then be assessed at the construction site level, with the optical sensor 10 being deployed in a stationary manner.

[0108] The system may further include storage means 30 configured to store measurements taken by the optical sensor 10. The storage means 30 may, alternatively or in addition, be configured to store the lookup table and / or the training database. The storage means 30 may also be adapted to store alert data for a specified storage period.

[0109] Storage options 30 may include a dedicated electronic card, such as an SD storage card. A dedicated electronic card consumes little power, thus preserving system battery life.

[0110] The system may also include an alert unit 40 adapted to alert a user when exposure to particles of the given material exceeds a predetermined exposure threshold. The alert unit 40 may correspond to the processing unit 20.

[0111] The user alerted by the alert unit may be the user wearing the system, for example an operator or inspector working on a construction site, or another user, for example an operator's manager or a site control center.

[0112] The system may also include a transmission unit 50 adapted for transmitting alert data. The transmission unit 50 can be configured to transmit via a wireless connection.

[0113] The transmission unit 50 can in particular include a transmitter module enabling transmission via LPWAN technologies, in order to limit the energy consumed by the transmission and thus increase the autonomy of the system.

[0114] The system may further include a battery 60 adapted to provide power to the processing unit 20 and / or the alerting unit 40 and / or the transmission unit 50.

[0115] Battery 60 can include one or more non-rechargeable batteries mounted in series, or be rechargeable by a charger, the type of battery being able to be chosen according to the desired autonomy for the system.

[0116] The system may further include a housing 100. The optical sensor 10 and the processing unit 20 are integrated into the housing 100. If necessary, the storage means 30, the alerting unit 40, the transmission unit 50 and / or the battery 60 may also be integrated into the housing 100.

[0117] The 100 housing can be made of waterproof material to protect the system's electronics from common external disturbances found on construction sites, such as water and dust. This increases measurement reliability and improves the lifespan of the device's components.

[0118] The 100 case can be portable, so that it can be carried by a user without hindering them, the 100 case having moderate weight and size.

[0119] Other embodiments may be considered and a person skilled in the art can easily modify the embodiments or examples set out above or consider others while remaining within the scope of the invention.

Claims

1. A method for determining an amount of particles of a given material in an air sample likely to include suspended particles of the given material and other materials, wherein said determination is carried out in real time simultaneously with an exposure of an individual to said air, and continuously, the determination being carried out periodically over time, in which the method comprises steps of: E1: optically measuring a signature of the sample by means of an optical counter configured to perform particle counting, wherein said sample signature is a function of a size and number of particles present in the sample and / or a shape and number of particles present in the sample, and wherein said sample signature comprises a histogram of the distribution of the number of particles in the sample as a function of particle size and / or a histogram of the distribution of the number of particles in the sample as a function of particle shape; E2: identifying, by means of a processing unit (20), a type of the sample by means of the measured sample signature and by means of : - either a classification model of the k nearest neighbor type and a training database comprising a plurality of reference signatures, each reference signature being associated with a type of a reference sample, identification of the sample type comprising a comparison of the sample signature with the plurality of reference signatures stored in the training database ; - or a classification model of the neural network, support vector machine, stochastic gradient algorithm or random forest algorithm type, said classification model being trained on a training database comprising a plurality of reference signatures, each reference signature being associated with a type of a reference sample; and E3: determining, by means of the processing unit (20) the amount of particles of the given material in the sample based on the identified sample type and a correspondence table associating, with each of a plurality of reference sample types, a reference amount of particles of the given material.

2. The method for determining an amount of particles according to claim 1, wherein the given material is at least one of respirable crystalline silica, ballast dust, lime, diesel particles, polluted earth dust, other hydraulic binders, Chromium VI particles, and / or other particles emitted by various site operations.

3. The method for determining an amount of particles according to claim 1 or claim 2, wherein the signature of the sample measured in step E1 depends on a number of particles present in the sample, and on a dimension of the particles present in the sample in the direction of the greatest dimension of particles of the greatest dimensions of particles present in the sample in two directions or in three orthogonal directions forming a three-dimensional space, and wherein the signature of the sample measured in step E1 comprises a histogram of distribution of a number of particles in the sample as a function of the dimension or one of the dimensions of the particles.

4. The method for determining an amount of particles according to claim 3, wherein the shape of the particles is determined as a function of the dimension(s) of the particles measured in the sample signature.

5. The method for determining an amount of particles according to any one of the preceding claims, wherein the signature of the sample measured in step E1 further depends on a color of the particles present in the sample and / or diffraction properties of the particles present in the sample, and wherein the signature of the sample measured in step E1 comprises a histogram of distribution of a number of particles in the sample according to the color of the particles and / or a histogram of distribution of a number of particles in the sample according to the diffraction properties of the particles.

6. The method for determining an amount of particles according to one of the preceding claims, wherein the sample type identified in step E2 is defined by a nature of the sample, corresponding to a material of the reference sample, and an activity of producing the sample, corresponding to an activity following which the reference sample was produced and / or to a tool corresponding to said activity.

7. The method for determining an amount of particles according to one of the preceding claims, wherein the signature of the sample is measured periodically in step E1 at each measurement period, the type of the sample is periodically identified in step E2 at each calculation period and the amount of particles is determined periodically in step E3 at each calculation period, wherein the calculation period is equal to or greater than the measurement period, wherein the measurement period and the calculation period are comprised between 1 second and 1 hour.

8. The method for determining an amount of particles according to one of the preceding claims, further comprising a step E4 of determining, by means of the processing unit (20) an exposure to particles of the given material from the amount of particles of the given material determined in step E3 and a time of exposure to said particles of the given material.

9. The method for determining an amount of particles according to claim 8, further comprising a step E5 of alerting a user when the exposure to the particles of the given material exceeds a predetermined exposure threshold.

10. A system for determining an amount of particles of a given material in an air sample likely to include suspended particles of the given material and other materials, said determination system, said system for determining being adapted to carry out a determination in real time simultaneously with an exposure of an individual to the particles of the given material, and continuously, the determination being carried out periodically over time, and in which said system comprises: - an optical counter (10) configured to perform particle counting adapted and measure a signature of a sample, said sample signature being a function of a size and number of particles present in the sample and / or a shape and number of particles present in the sample, and said sample signature comprising a histogram of the distribution of the number of particles in the sample as a function of particle size and / or a histogram of the distribution of the number of particles in the sample as a function of particle shape; and - a processing unit (20) configured to: - identify a type of the sample by means of the measured sample signature and by means of: - either a classification model of the k nearest neighbor type and a training database comprising a plurality of reference signatures, each reference signature being associated with a type of a reference sample, identification of the sample type comprising a comparison of the sample signature with the plurality of reference signatures stored in the training database; - or a classification model of the neural network, support vector machine, stochastic gradient algorithm or random forest algorithm type, said classification model being trained on a training database comprising a plurality of reference signatures, each reference signature being associated with a type of a reference sample; and - determine the amount of particles of the given material in the sample based on the identified sample type and a correspondence table associating, with each of a plurality of reference sample types, a reference amount of particles of the given material.

11. The system for determining an amount of particles according to claim 10, wherein the optical counter (10) is portable.

12. The system for determining an amount of particles according to claim 10 of claim 11, further comprising storage means (30) configured to store the measurements made by the optical counter (10) and / or the training database and / or the correspondence table.

Images